Autism spectrum disorders are now among the most prevalent medical conditions of childhood. Only a small fraction of the 486,000 individuals under 20 years of age with an autism-spectrum disorder (ASD) in the U.S are young enough to benefit from intensive early intervention. Overall prognosis for the older children with autism is not good. Despite improvements in treatment and education over the past 30 years, adult outcome even for non-mentally retarded individuals with autism has not significantly improved. The majority continue to need high levels of parental and community support throughout their lives. One reason for this huge public health problem is that the brain-basis of fundamental deficits in older children and adults with autism is not well understood. We propose collaboration between a longitudinal neuroimaging, clinical, and neuropsychological study of late neurodevelopment in autism and the National Alliance for Medical Imaging Computing (NA- MIC), one of the NIH Roadmap National Centers for Biomedical Computing. The collaboration will bring state-of-the-art brain imaging analysis tools developed by NA-MIC into autism clinical research and form a new highly interactive multidisciplinary research team. Working together, the computer scientists and clinical researchers will use critical biological questions in autism to "drive" the development of NA-MIC "tools". The critical biological questions are 1) what is the microstructural basis of abnormal brain connectivity during late neurodevelopment in autism, and 2) how is brain microstructure related to deficits, developmental trajectory, and outcome. We will use Time 1 and Time 2 high-resolution MRI and diffusion tensor imaging data that have already been collected on a single 3Tesla scanner on a cohort of 100 males with high-functioning autism and 72 typically developing males. Time 3 and 4 data are being collected over the next 5 years (MH080826). We will use novel, non-tractography-based, diffusion tensor image analysis methods developed by NA-MIC to compare, at the level of both individuals and groups, microstructural features along entire white matter tracts in language, social, and repetitive behavior neural networks. We will integrate the white matter analysis with structural image analysis of cortical and subcortical gray matter. We will determine how microstructural white matter features, gray matter morphometric features, and clinical deficits are related to each other and change over time in autism. Autism is now one of the most prevalent medical disorders of childhood. Adult outcome for even high-functioning older children and adolescents with autism is still usually poor, with a high need for lifelong clinical and community support. Understanding the brain- basis of deficits, disease course, and outcome in these individuals is critical for development of autism-specific treatments as well as secondary and tertiary preventive interventions.